Skip to content

Translation in progress Some entries still include text in the original language.

EventPast event

Alice Labs发布Cyber RL Benchmark v1.0

1 reports1 reporting sourcesLatest report

Start here

AI overview

Alice Labs 发布 Cyber RL Benchmark v1.0,在 12 项真实软件漏洞任务上评测 10 款前沿模型,覆盖恶意软件分析、漏洞检测、漏洞修补与黑盒渗透测试四类。每任务每模型跑 3 次,按 0-1 分部分给分,被护栏拦截计 0 分。Qwen 3.8 Max 以 0.72 均分、解出 7 项任务居首,Mistral Medium 3.5 以 0.36 垫底。漏洞修补类最高有 6 款模型并列 100%。

Generated by AI from reports · Updated 3 days ago

Progress timeline

Oct 7
  1. Alice Labs
    Cyber RL Benchmark v1.0 发布:评测 10 款前沿模型的网络安全能力

    Alice Labs 发布 Cyber RL Benchmark v1.0,在 12 项真实软件漏洞任务上评测 10 款前沿模型,Qwen 3.8 Max 以 0.72 均分、解出 7 项任务居首,Mistral Medium 3.5 以 0.36 垫底。基准覆盖恶意软件分析、漏洞检测、漏洞修补与黑盒渗透测试四类,每任务每模型跑 3 次并按 0-1 分部分给分,被护栏拦截计 0 分。漏洞修补类最高有 6 款模型并列 100%,黑盒渗透测试最难,最高仅 42%;Opus 5.5 的护栏在 API 层拒绝了全部渗透测试任务。

Related discussion

No discussion links have been collected for this event yet.

Attention trend

New sources per day

6 days total1 ·Maximum1(Oct 7)·Today0

  • Oct 7New1 sources (1 outlets,0 accounts)
  • Oct 8New0 sources (0 outlets,0 accounts)
  • Oct 9New0 sources (0 outlets,0 accounts)
  • Oct 10New0 sources (0 outlets,0 accounts)
  • Oct 11New0 sources (0 outlets,0 accounts)
  • Oct 12New0 sources (0 outlets,0 accounts)

Hourly attention

Current attention31·Peak in the comparable period35(Oct 7 17:00)·Change over the comparable past 24 hours 0%

010203040Oct7Oct9Oct11Oct12